China AI chip firm Biren raises new funds, plans Hong Kong IPO
Chinese AI chip startup Biren Technology has raised about 1.5 billion yuan ($207 million) in fresh funding and is preparing for a Hong Kong initial public offering, people familiar with the matter said.
ADVERTISEMENT The funding round and IPO plan come as China seeks to develop domestic alternatives to U.S. semiconductors amid escalating export restrictions by Washington on advanced chips.
Beijing has prioritised building homegrown champions in graphics processing units (GPUs), which are critical for artificial intelligence development.
The 1.5 billion yuan funding round was led primarily by state-linked investors, two of the sources said. Participants included a state-backed fund from Guangdong province and another from the Shanghai government, according to one source. Biren initially filed documents for a mainland China listing last year but has since shifted focus to Hong Kong, partly due to stricter regulatory requirements on the mainland including less tolerance for loss-making companies, two of the sources said. The company is preparing to file for a Hong Kong listing in the third quarter, potentially as early as August, one source said. It was not immediately clear if Biren had appointed advisers for the IPO.
ADVERTISEMENT Biren was valued at approximately 14 billion yuan prior to the latest funding round, two of the sources said. Biren, and the Guangdong provincial and Shanghai governments, did not immediately respond to requests for comment.
ADVERTISEMENT The sources, who spoke to Reuters, declined to be named as the information is not public.
Developing domestic GPU capabilities has become critical for China as the U.S. has tightened export restrictions on advanced semiconductors. The latest measures, implemented in April, prompted U.S. chip giant Nvidia to halt sales of its H20 AI chips to Chinese customers.
ADVERTISEMENT
The potential market for Chinese AI chip companies is substantial. Investment bank Morgan Stanley predicted in a May client note that domestic GPU makers could generate 287 billion yuan in sales by 2027, capturing 70% of the Chinese market compared with 30% last year.
Faced upheaval
Founded in 2019, Biren's cofounders include Zhang Wen, formerly a president at leading AI face-recognition company SenseTime, and Jiao Guofang, who previously worked at Qualcomm and Huawei.
ADVERTISEMENT The company initially drew attention from investors and industry observers in 2022 when it unveiled its first batch of products, including the BR100 chip, which it claimed could match the performance of Nvidia's advanced H100 AI processor. However, the company was added to the U.S. 'Entity List' in 2023, preventing it from using leading global foundries such as Taiwan Semiconductor Manufacturing to manufacture its chips. Biren has since experienced significant upheaval, with some senior executives departing, including co-founder Xu Lingjie. It continues to operate at a loss and generates limited revenue, recording 400 million yuan in sales in 2024, according to two of the sources.
The company's general-purpose GPU products have been deployed across multiple intelligent computing centers, and its partners include China Mobile, China Telecom, ZTE, and the Shanghai AI Laboratory, according to its official website. Some of these companies have also been targeted by U.S. restrictions, including China Mobile and China Telecom, which the Federal Communications Commission said in March it is investigating for potential evasion of U.S. sanctions.
Biren also faces intense competition from other Chinese AI chip companies, including Huawei and peers such as Tencent-backed Enflame and Metax.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Mint
an hour ago
- Mint
India's rare earths riddle: No quick fix as China holds the magnetic key
There seems to be no quick fix in sight for India's rare earths predicament as it scrambles for magnets, slaps export curbs and weighs production incentives, while a delegation from the country waits to meet Chinese officials. Issues related to raw material availability, global magnet supply, import duty structure and access to technology are key constraints that could delay any local manufacture of rare earth magnets in the short to medium term, said S.B. Mohanty, acting chairman and managing director of IREL Ltd, the state-owned rare earths company. On 4 April, China restricted exports of seven rare earth elements and magnets critical in defence and automobile sectors. Since Indian automakers have relied exclusively on China for magnets, the crunch threatened production at companies like Tata Motors Ltd and Maruti Suzuki India Ltd, triggering a search for alternatives. "We don't make the more powerful batteries that automakers want, because we don't have access to a few rare earth elements that are crucial raw materials in the manufacturing process," Mohanty said in an interview. IREL mines and refines critical raw material that goes into rare earth magnets; and at full capacity, it could cater to a third of India's requirement. However, its plants operate at half capacity, and the company doesn't make the magnets themselves. On 17 June, Mint reported that the company does not have a full-time CMD since December. Currently, China controls 90% of the world's battery production and Japan about 7%, while the rest come from countries like Vietnam and Thailand which account for 1% each. Japanese firms make advanced batteries with specialized applications, some of which even China imports. That essentially means for any immediate requirement, India has no option but to lean on China. "If China relaxes its restrictions, then we will return to normal; but if it doesn't, there are no quick solutions," Mohanty said. On 13 June, Reuters reported that India has asked IREL to suspend a 13-year-old agreement on rare earth exports to Japan to save supplies for domestic needs. At a recent meeting with industry executives, commerce minister Piyush Goyal asked IREL to stop exports of rare earths, mainly neodymium, used in magnets for EV motors, the report added. Union industry minister H.D. Kumaraswamy last week said the government will decide on a scheme to support domestic production of rare earth magnets in 15-20 days. Separately, India is looking to source rare earth magnets from countries such as Vietnam and Japan, Mint reported on 24 June, citing a government official. The government is currently talking to industry stakeholders on the quantum of government support to process rare earths in India. According to IREL's Mohanty, even though a factory to make auto-grade rare earth magnets can be built in a year for ₹800 crore with some help from the government, the problem will be in sourcing the raw materials. Among the seven elements in the Chinese ban list are yttrium, gadolinium and lutetium, which enhance the performance of a normal rare earth battery at high temperatures. India has few deposits of these elements and China imports them from Myanmar, Vietnam and Sri Lanka. "It is not as if we don't have these metals, but our mine yields are so low that they are unremunerative and that will require us to look for new sources of imports like Australia or even Myanmar," says Mohanty. "We will also have to find out who will sell us equipment for factories as they can't be ordered off the shelf." It's not that there was no attempt to make rare earth magnets domestically. As the industry's nodal agency, IREL had invited expressions of interest from private players to set up manufacturing units in 2020. "A few private players showed interest, but all of them backed out as the import duty structure for raw materials was more than intermediate and finished products, and that made the project unviable. At that time, the lower duty on intermediates was essentially to favour user industries like automobiles," says Mohanty. For the future, Mohanty has a few recommendations. He says that for the short term, India should stockpile batteries, increase capacity utilization (which IREL is doing now), source and license global technology, and recycle more to create a circular economy. He says that while India recycles only 5% of its old magnets, Chinese rates stand at over 80%. "We must understand that the key levers in this business are with China. Our batteries are subsidized to the extent 35% already to match Chinese prices, and if we do manage to increase production, China can drop prices more and make our new factories unremunerative," says Mohanty. "But, if we want to secure our industries in the long run, there is no option but to create local alternatives whatever the cost."
Mint
an hour ago
- Mint
How much energy does your AI prompt use? I went to a data center to find out.
There they were. Golden boxes, louder than a toddler on a red-eye, hotter than a campfire in a heat wave, pricier than a private Caribbean island. Yes, real, working Nvidia GPUs. I was under strict 'look, don't touch" orders—as if I'd, what, lick the mesh metal enclosure. Just standing there, I could hear and feel the electricity being devoured. We've all heard about AI's insatiable energy appetite. By 2028, data centers like this one I visited in Ashburn, Va., could consume up to 12% of all U.S. electricity, according to a report from the Energy Department and Lawrence Berkeley National Lab. And yes, we're the problem, it's us. (Insert Taylor Swift-related groan here.) Every time we ask AI to write an email, draw an anime-style George Washington or generate a video of a cat doing a back flip, we're triggering another roar in those massive halls of GPUs. What I wanted to know was, how much power do my AI tasks actually use? The equivalent of charging a phone? A laptop? Cooking a steak on an electric grill? Powering my house? After digging into the research, visiting a data center, bugging just about every major AI company and, yes, firing up that grill, I got some answers. But not enough. Tech companies need to tell us more about the energy they're using on our behalf. Let's start with a recent, popular example: 'a video of a cat diving off an Olympic diving board." The moment you hit enter, that prompt gets routed to a massive data center. When it arrives, it kicks off inference, where pretrained AI models interpret and respond to your request. In most cases, rows of powerful Nvidia graphics processing units get to work turning your weird idea into a weirder reality. Rival chips from companies like Amazon, Google or Groq are also starting to be used for inference. The model training itself happens earlier, with Nvidia chips. The facility where I saw that 'SuperPod" of Nvidia H100 GPUs was run by Equinix, one of the world's largest operators of data centers that provide cloud infrastructure—and now, AI. Chris Kimm, Equinix senior vice president of customer success, said that while AI training can happen just about anywhere, inference is best done geographically closer to users to deliver the best speed and efficiency. Figuring out how much energy your individual AI prompts use would be a lot easier if the major AI companies actually shared the darn info. Google, Microsoft and Meta declined. Google and Meta pointed me to their sustainability reports. OpenAI shared something. Chief Executive Sam Altman said that the average ChatGPT query uses about 0.34 watt-hours of energy. OpenAI wouldn't break out details on text, image or video energy usage. Researchers have stepped in to fill the gap. Sasha Luccioni, the AI and climate lead at open-source AI platform Hugging Face, has run tests to estimate the energy required to generate different types of content. Along with other researchers, she also maintains an AI Energy Score leaderboard. Since the top AI players use their own proprietary models, she relies on open-source alternatives. The energy required to generate content varies widely depending on the model and GPU setup. Compare Luccioni's findings with charging a typical smartphone, which uses around 10 watt-hours of energy: • Text: A lightweight, single-GPU Llama model from Meta used about 0.17 watt-hours, while a larger Llama model running across multiple GPUs used 1.7 watt-hours. • Images: Generating a single 1024 x 1024 image with one GPU also used 1.7 watt-hours. • Video: This is the most intensive. Even making 6-second, standard-definition videos used anywhere between 20 and 110 watt-hours. I wanted to better understand the stakes—literally. So I grabbed an electric grill from Home Depot, a power meter and my video producer, David Hall. About 10 minutes and 220 watt-hours later, we had a thin, medium-well steak. Translation: The energy it took to cook a decent dinner was about the same as generating two AI videos, at the high end. (Watch the video above for more steak breakdowns.) Remember the short AI film I made using Google Veo and Runway a few weeks ago? We generated about a thousand 8-second, 720p clips for our film. Going by these estimates, we might have used roughly 110,000 watt-hours. That's nearly 500 steaks! But, as I said, Luccioni doesn't have the power-consumption data for the commercial AI tools, and her numbers aren't a perfect match: On the one hand, our video was higher quality than the 6-second, 480p clips in Luccioni's research. On the other hand, the popular video models are likely optimized for greater efficiency, experts say. 'Until we get access to these models," Luccioni said, 'all we can do is estimate." Her tests also use Nvidia's last-generation Hopper chips. Nvidia has seen a jump in energy efficiency with its latest Blackwell Ultra chips, according to Josh Parker, the company's head of sustainability. 'We're using 1/30th of the energy for the same inference workloads that we were just a year ago," Parker said. That said, plenty are still using those older chips. The pod I saw at Equinix's facility? It cost over $9 million in just Nvidia hardware alone. You don't just toss that in the dumpster when new ones come out. And I've only covered electricity. These hot GPUs also require a lot of water to stay cool, but that's a whole other story. Data-center providers and tech companies I spoke to all said the same thing: Demand for these GPU-filled buildings keeps multiplying. Just driving through Ashburn, I saw five massive data centers going up. The companies also stressed the improving efficiency of models and chips, and their efforts to shift to cleaner, more renewable energy sources. No matter how efficient things get, more of us are using AI. We could all buy more efficient air conditioners, but if the planet keeps getting hotter, we're going to crank the AC more—and burn more energy. Luccioni hopes we at least consider energy use when we use these tools, maybe think twice about generating a dozen cat videos. And it's on the companies to start sharing real numbers, so that we can make informed choices. Back to Virginia, and those screaming GPUs. Turns out, they weren't generating Olympic kitty videos. They were owned by Bristol Myers Squibb—and they were searching for new cures to diseases. Not all AI prompts are what you'd call a waste of energy.
India.com
2 hours ago
- India.com
Not Iran, The Real Threat Is Pakistan… Tensions Rise After US Report, How India's Warning Is Proving True
New Delhi: The maps in Washington have shifted focus. The chatter inside intelligence rooms is not about Tehran anymore. It is Rawalpindi. A fresh report from U.S. intelligence agencies has triggered serious alarm. Hidden under radar, Pakistan is reportedly building an intercontinental ballistic missile. It is said to be nuke-tipped that could reach mainland America. Until now, Iran had held the spotlight. But this new revelation has brought Pakistan's nuclear ambitions straight into global crosshairs. For India, it is vindication. For the United States, a potential crisis. For the region, a ticking clock. A report quoting U.S. sources claims that Pakistan's new missile has crossed theory stage. It is being developed quietly – away from public eyes. It is capable of traveling over 5,500 kilometers. That is ICBM territory. A range wide enough to strike far beyond Delhi. All the way to Washington. The report was published in Foreign Affairs – an American magazine. It says the missile project picked up speed after India's Operation Sindoor rattled Pakistan's military brass. With Chinese help, Pakistan is now said to be acquiring the components and know-how needed to leapfrog its existing arsenal. What This Means for America Washington is watching closely. Senior officials say that if Pakistan achieves the ICBM capability, it will be reclassified as a nuclear threat like North Korea, China and Russia. There will be no middle ground. The United States has already imposed sanctions. Last year, it froze assets and blocked trade with Pakistan's National Development Complex and its key missile contractors. That was not random. It came after evidence emerged of Islamabad sourcing sensitive tech from foreign markets. For years, Pakistan insisted its nuclear weapons were for deterrence against India. Nothing more. No ICBMs. No aggression. Just defense. But this new intelligence contradicts that script. With Shahin-III already reaching over 2,700 km, experts believe Pakistan is preparing to extend its reach far beyond South Asia. And there is a motive. If Pakistan owns a missile that can hit the United States, it will change the rules. It will act as a shield. It will limit U.S. involvement in any future India-Pakistan conflict. That is the strategy. Quietly but clearly. Why India Is Saying 'We Told You So' New Delhi has long raised concerns. That Pakistan cannot be trusted with nuclear weapons. That its command chain is fragile. That terror groups work hand in glove with rogue military elements. After recent air strikes inside Pakistan, India also neutralised nine terror camps and targeted 11 airbases. Pakistan reportedly fired a Fateh-II hypersonic missile. India shot it down. Defence Minister Rajnath Singh and other top Indian leaders have called on the International Atomic Energy Agency (IAEA) to monitor Pakistan's nuclear stockpiles. The concern is not hypothetical. It is urgent. One weapon in the wrong hands could mean catastrophe. While the world debated strikes on Iran, Pakistan was quietly reshaping its arsenal. The fear is not merely the missile. It is what surrounds it – terror cells, sleeper agents and political instability. If these missiles fall into the wrong hands, no country is out of reach. No continent immune. According to global estimates, Pakistan currently has around 170 nuclear warheads. It is not a signatory to the Nuclear Non-Proliferation Treaty. It also has a history of proliferation. That is what makes the new report even more chilling. The ICBM threat from Pakistan is no longer whispers. It is a documented risk. A weapon in the making. America now faces a choice. Ignore and risk a future crisis. Or act. India, meanwhile, continues to say the same thing - Pakistan's nukes are not only a South Asia's problem. It could soon become the world's.
